Metal complexes of azo dyes. I. Quadridentate complexes from

water quality after wildfires. The Camp Fire is the deadliest and most destructive wildfire in California's history. The fire killed... BUSINES...
28 downloads 0 Views 975KB Size
It can be applied to any type of

Listings and/or loans of the decks for these programs will be sent on request.

rapid and reliable. structure.

Conclusions The adaptation of the Bailar method of isomer enumeration to the computer has been shown to be

Acknowledgments.-We are grateful to the University of Iowa Computer Center and the Graduate College for computer time.

RESEARCH AND DEVELOPMEST DIVISION,'" LABORATORY, ORGANICCHEMICALS DEPARTMEST, C O X P A N Y , ISC., mTILMINGTON, DELAWARE19899

CONTRIBUTION FROM THE JACKSON

E. I. DU P O N T

DE

NEMOURS AND

Metal Complexes of Azo Dyes. I. Quadridentate Complexes from Bidentate Azo Compounds and Alkanediamines or Ethanolaminelb BY THEODOR A. LISS A N D DOXALD R. BAER

Received Noovenzber 21, 1968

o-Halo-0'-hydroxydiarylazocompounds react readily with alkanediamines or ethanolamine in the presence of copper(I1) or nickel(I1) salts to form in high yield the quadridentate complex of the ligand resulting from nucleophilic displacement of the halogen by the amine. The order of reactivity is I > Br > C1 > F, but even the fluoro compound reacts a t room temperature. In polyhaloazo compounds only the ortho halogen is reactive under the reaction conditions. 4-(o-Bromophenylazo)3-methyl-l-phenyl-2-pyrazolin-5-one ( X I ) reacts with ethylenediamine, alone or in the presence of Cu(II), to give the uncomplexed ligand XII, which can be metalized in a separate step with cupric chloride in ethanolamine. The fluoro and bromo compounds Ia and IC,respectively, also react with ethylenediamine in the absence of metallic salts to form the quadridentate ligand XIT'. The o-amino-0'-halo compounds (XVIIIc and XVIIId) react with ethylenediamine in a surprising manner to form the reduction product XX. The o-bromotriphenylformazan X X I reacts to form a complex of the four-nitrogen ligand XXII in which the copper is five-coordinate. The o-(methoxycarbonyl) azo compound XXIII reacts in one step to give the quadridentate complex XXIT'.

compounds obtained by removing the metal from the Metal complexes of azo compounds suitably subfirst-formed metal complexes, which themselves were stituted in the o,o' positions constitute an important not characterized. class of commercial dyes2 Most of these dyes are Three isolated reports of the use of bifunctional combased on azo compounds which are tridentate and which pounds in the metal-catalyzed reaction of o-ha1o-o'therefore form bicyclic metal complexes. Quadridenhydroxyazo compounds have been located. The patent tate and quinquidentate azo compounds which form triliterature6 reports the use of 8-aminoquinoline and and tetracyclic metal complexes have also been reformulates the products as quadridentate complexes. ported. The requisite multidentate azo compounds Delfs3b mentions the use of ethanolamine in a patent were first prepared and were then treated with a metal example. Finally, Bamberger and Gross' report the ion in a subsequent step. treatment of o-chloro-0'-hydroxyazo dyes on a fiber with The present work is concerned with the reaction of a copper compound and an ortho-substituted aromatic bidentate azo compounds with alkanediamines or amine or a heterocyclic amine. It was suggested that ethanolamine in the presence of copper or nickel salts to the tricyclic metal complex had formed on the fiber. form metal complexes of quadridentate azo compounds The alkanediamines and ethanolamine used in the in one step. present work have the advantage of being good solvents The replacement of the chloro group in o-ch1oro-o'for both the azo compounds and the metal salts and hydroxyazo compounds by hydroxy, alkoxy, alkylsimultaneously serve as a base to neutralize the acid amino, and arylamino, as well as other groups, in the generated in the reaction. Much of the earlier work on presence of cupric salts was first reported by Delfs and metal-catalyzed reactions of o-halo-0'-hydroxyazo comcoworkers. This reaction was subsequently extended was carried out in heterogeneous systems in and extensively studied by Stepanov and c o ~ o r k e r s . ~ , pounds ~ which two or even three phases were present. Both groups were primarily interested in the new azo (1) (a) Publication No. 419. (b) T h e majority of this work was included in a paper presented by D. R. Baer a t the Third Internationales Farbensymposium, Interlaken, Switzerland, M a y 18, 1967. (2) For a comprehensive review of metal complexes of dyes, see H. Baumann and H. K. Hensel, Fortschr. C h e m Foisch., 7, 643 (1967). (3) (a) 1).Delfs and K. Knoche, German P a t e n t 571,859 (Sept 29, 1931); (b) D. Delfs, German Patent 658,831 (March 18, 1934); (c) D. Delfs, German P a t e n t 738,900 (Oct 31, 1936).

(4) F o r paper X X I I in the series, see B. I. Stepanov and M. A. Andreeva, Souiet J . O i g . Chem., 2 , 2171 (1966). (5) A review of this work is given by B. I. Stepanov in "Recent Progress

I n the Chemistry of I i a t u r a l and Synthetic Colouring Matters," Academic Press, New York, S . Y., 1962, pp 461-471. ( 6 ) German P a t e n t 748,913 (July 25, 1939). (7) R. Bamberger and R. Gross, Teztil-P~axis,16, 931 (1961); 17, 485 (1963).

METALCOMPLEXES OF Azo DYES 132'3

Vol. 8,No. 6 , June 1969

o-Halo-0'-hydroxyazobenzenes When a solution of the yellow o-chloro-0'-hydroxyazo compound I b was heated to 85-95" for several hours with 1 equiv of cupric chloride in aqueous ethylenediamine solution, a marked bathochromic shift was observed and complex I1 was obtained in 92y0yield. The reaction also occurred a t room temperature if the reaction time was lengthened. ?HI

m

+

HJ'!CH?CH>NH?

cu'+

I

x

HO

Ia, X = F

x

b, = c1 c,X=Br d,X=I

I1

The structural assignment8 is made on the basis of the elemental analysis, method of synthesis, and the infrared and visible absorption spectra. The infrared spectrum of I1 shows a doublet of weak to moderate intensity a t 3180 and 3110 cm-l (N-H stretch), which confirms the presence of a coordinated primary amino group. The infrared spectra of coordination compounds of ethylenediamine, for example, generally show doublets or even multiplets in the N-H stretching region of about 30503350 cm-l; thus, Cu(en)zPtC14 absorbs a t 3320 and 3245 cm-la9 When the bromonaphthol I11 (TableI) was used as the halogen component, the naphthalene analog (IV, Table 11) of I1 was obtained. Substitution of 1,2-propanediamine for the ethylenediamine resulted in formation of complex V (Table 11). Two isomers can form in this reaction as a consequence of the dissymmetry of the diamine; thin layer chromatograDhy of the product indicated that a mixture of the two isomers was obtained. Similarly, 1,3-propanediamine reacted with bromo compound ICand cupric chloride to form complex VI which has two six-membered rings as well as a fivemembered ring surrounding the metal atom. The

absorption maximum occurs a t 566 nm (DMF), which is virtually the same as that for the analog from ethylenediamine (567 nm in DMF). The infrared spectrum of VI showed a doublet of weak to moderate intensity a t 3230 and 3110 cm-' (N-H stretch). As indicated above, both o-chloro- and o-bromophenylazo compounds react with alkanediamines to form copper complexes. The o-fluoro compound I a was found to be somewhat less reactive but nevertheless reacted with ethylenediamine a t 70-SO O within several hours to give a good yield of 11. The reaction even occurred slowly a t room temperature; after 7 months a 70% yield of I1 was isolated (this long reaction time was probably not necessary). The reactivity of the fluorine in the present system contrasts with the report that no reaction was observed between 1-(0fluorophenylazo)-2-naphtholand sodium methylate in the presence of copper acetate after 8 hr a t 100" (methanol-dioxane solvent). lo The iodo compound Id, as expected, was the most reactive compound in the series. Although no quantitative rate data were obtained, Id was observed to react readily a t room temperature. When polychloro azo compounds were allowed to react with ethylenediamine in the presence of cupric salts, only the o-chlorine was reactive under the conditions used. Thus, the trichloro compounds VI1 and I X yielded complexes VI11 and X.

OH VII, X = H. Y = C1 IX, X = C1, Y = H

Y

(8) The copper is shown coordinated t o the azo nitrogen attached to the p-cresol ring. This assignment is arbitrary. Isomer IIa is also possible. To our knowledge there has been no report of isolation or detection of isomers of this type.

VIII, X = H, Y = CI X, X = C1, Y = H

A surprising result was found in the reaction of the pyrazolinone X I with ethylenediamine. Both in the L N H > IIa (9) D. B. Powell and N. Sheppard, Spectvochim. Acta, 17, 68 (1961).

(10) M. A. Andreeva and B. I. Stepanov, J . Gen. Chem. U S S R , 30, 2750 (1960).

1norgcrnic Chemistry

1330 THEODOR A. LISS A N D DOICALD R.BAER hl

d

? m

Y

Lo

01

hl

'9

4

'9

M

?

M

r

? l

r l r i

w

14

M

M

METALCOMPLEXES OF A z o DYES 2,731

Vol, 8, No. 6, June 1969 presence and in the absence of cupric ion, the product was the uncomplexed quadridentate compound X I I . l 1 I-\

0 XI

CH, I

I-\

solutions of XI11 are deep blue. Thus, in acetone and dimethylformamide the maxima appear a t 590 ( E 8100) and 598 nm ( e 8100), respectively. The above results with the pyrazolinone compound suggested that the reaction in the azobenzene series might also occur in the absence of cupric ion. I n fact, it was found that both the fluoro and bromo compounds I a and IC,respectively, react with ethylenediamine a t 75-85' to form the quadridentate compound XIV. I n a subsequent step XIV was treated with cupric sulfate in pyridine to form complex 11, identical with material obtained as described earlier.

pd

CH,CH,NH,

\ /

XI1 I-\

I

OH

&H,CH,NH, XIV

XI11

The infrared spectrum of X I 1 (KBr pellet) shows a sharp band of weak to moderate intensity a t 3390 cm-I and a broad complex base a t 3100-2400 cm-l. The absence of the doublet characteristic of a primary amino group (N-H stretch) is probably due to bonded interactions (see below). Evidently the cupric ion is more tightly complexed by the ethylenediamine than by X I I , so that under the reaction conditions the complex of XI1 is not formed. The copper complex XI11 was obtained by allowing X I 1 to react with cupric chloride in aqueous ethanolamine solution. Ethanolamine forms a complex with cupric ion much less strongly than does ethylenediamine and therefore does not prevent formation of X I I . The use of ethanolamine as solvent is rather specific; in solvents such as methanol and pyridine only impure copper complex was obtained. The infrared spectrum of XI11 shows a multiplet in the N-H stretch region, with a very weak band a t 3380 cm-' and weak to moderate bands a t 3250 and 3200 cm-l. The electronic spectral properties of complex XI11 bear mentioning. The quadridentate compound X I 1 has a maximum a t 425 nm (E 16,000) in methanol, whereas the brown copper complex XI11 gives a yellow solution in methanol with a strong maximum a t 415 nm (E 19,800). This hypsochromic shift on metalization contrasts with the bathochromic shift observed in the analogous azobenzene derivatives X I V and I1 (see below).12 I n addition to the maximum a t 415 nm ( E 19,800), complex XI11 has a broad shoulder of low intensity beyond 500 nm (E -1800). I n other solvents, this shoulder absorption becomes a well-defined peak a t considerably longer wavelengths, with the result that (11) T h e o-chloro analog of XI was reported t o be unreacted with sodium butoxide and copper acetate i n a mixture of 1-butanol and pyridine, even after 8 hr at 100": M. A. Andreeva and B. I. Stepanov, J.Gen. Chem. U S S R , SO, 1380 (1960).

The infrared spectrum of XIV taken as a Nujol mull or KBr disk showed only a broad and complex band of weak to medium intensity a t 3100-2400 ern-', with no absorption a t higher frequencies. A carbon tetrachloride solution of XIV, however, showed a weak, moderately sharp doublet a t 3380 and 3240 cm-l expected for a primary amino group, in addition to a weak, broad band from 3100 to 2400 cm-'. Thus, even in dilute solution, bonded interactions still occur in XIV; these interactions presumably represent intramolecular hydrogen bonding to the azo group. The bromo compound IC reacts with ethanolamine and cupric chloride a t room temperature to give the complex XV in high yield. No reaction was observed a t room temperature in the absence of cupric chloride. Ethanolamine could presumably react to replace the bromine with either the hydroxy or the amino end of the molecule. Since XV shows an absorption maximum a t 566 nm (dimethylformamide), which is virtually the (12) However, a similar hypsochromic shift is observed with pyrazolinones in the tridentate series (as shown in t h e equation below), whereas again a bathochromic shift occurs on dealkylative metalization of o-hydr0xy-o'methoxyazobenzenes; unpublished observations.

max 405

11111

1332 THEODOR A. LISSAND DONALD R. BAER

Inorpitic Chemistry TABLE I1

PHYSICAL PROPERTIES AND ANALYSES OF METALCOMPLEXES

p - Iq cu-0 I B

A-

R Compd

Paient azo compd

A-R-B

Mp, 'C

263 dec 209-210 dec

I1 IV

Ia-d

I11

S-(CHs)?-NHa N-(CHd-iTHz

v

IC

N-CH-CH-NHn

I

1

A:p, nm5

e

567 573 452 568

22,600 24,800 7,200 21,700

566 587 476 571 598C 630 582 509 566 503

7--70 c---

Empirical formula

---70

7--70

H---.

h----

Calcd

Found

Calcd

Found

Calcd

Found

CxHiaS40Cu CisHiaN10Cu

54.3 58.8

54.6 58.8

4.86 4.39

4.8 4.2

16.9 15.2

16.9 15.1

CiaHiaNiOCu

55.6

55.5

5.24

5.2

16.2

16.3

20,200 21,300 5,000 23,100

CieHisNaOCu C14Hi2ClzNaOCu

55.6 43.5

55.6 43.9

5.24 3.13

5.3 3.3

16.2

14.5

16.1 14.4

43.5 CI~HIL!~~S~OCU

43.6

3.13

3.2

14.5

14.5

8,100 6,500 6,000 16,200 22,100 4,200

CisHisS~OCu CuHlaNaONi

54.3 55.1

54.0 55.3

4.56 4.93

4.6 5.2

21.1 17.1

20.8 17.2

CiaHiaNaOnCu C~sHiaKaOzCu

54.1 53.4

54.2 53.4

4.55 4.48

4.4 4.6

12.6 15.6

12.4 15.4

L-J VI VI11

IC \'I1

H,CHa S-(CHx)a--NHz h--(CH2)2-hTH2

X

IX

N-(CHz)2--SHz

XI11 XVIb

XI IC

S-(CHs)r-SHz hT-(CHs)-NH2

xv

IC XXIII

N-(CH2)*-0H CN-(CHz)z-NHz

XXIV

I1

211-212 253-254 266-267 dec 197-200 296-300

>300 274-275 dec

0 a

Absorption peaks from 400 to 800 nrn are listed.

* Sickel complex.

same as for complex I1 (567 nm), i t is clear that it is the amino group which is attached to the benzene ring. This assignment is confirmed by the infrared spectrum of XV, which shows no absorption in the N-H stretching region but has a broad complex band of weak intensity in the 2900-2000-~m-~region. This shift of the 0-H stretching vibration to lower frequency, as well as the breadth of this absorption band, evidently reflects substantial bonded interaction in XV. A nickel complex XVI corresponding to I1 mas obtained by a procedure similar to that for the copper complexes, ie., reaction of IC with aqueous ethylenediamine and nickel chloride a t 86-95'. Elemental analysis indicates that the octahedral positions of the nickel are not occupied by external ligands. The electronic absorption spectrum of XVI shows welldefined peaks in methanol a t 618 nm (E 6300) and 573 nm ( E 5100)) with the strongest peak a t 482 nm (E 12,000). The analogous copper complex I1 shows only weak shoulder absorption beyond 600 nm and has its strongest peak a t 551 nm ( E 20,700). I n view of the facility of the above replacement reactions with amino compounds, an attempt was made to extend the reaction to the use of ammonia.13 This goal (13) In i e f 3b the use of ammonia is suggested, but no specific examples are given.

0

See text,

isyparticularly desirable because of the difficulty in preparing o-amino-0'-hydroxyazobenzenes(and therefore their complexes) by other routes. For the present reaction a solvent was needed which is compatible with the aqueous copper-ammonia complex and which will solubilize the 2 : 1 o-halo-0'-hydroxyazo dye-copper complex which forms as an intermediate. With bromo compound ICand cupric chloride in a mixture of glyme and ammonium hydroxide, very little reaction occurred (other than formation of the 2 : 1 complex, which is very soluble in the glyme) even up to 200'. The same result was observed with concentrated ammonium hydroxide in dimethyl sulfoxide a t 100". After reaction a t 170" in formamide it was possible to isolate a small amount of the benzotriazole XVII, as well as unidentified products. The formation of the benzotriazole XVII suggests that replacement of the

'x' HO XVII

o-bromo substituent by amino did occur, but that under the vigorous reaction conditions the o-aminoazo compound thus produced was oxidatively cyclized14 under the influence of the cupric ion. The reason for the much lower reactivity of ammonia in the replacement reaction is not clear. (14) poi- a discussion of the cyclization of a-aminoazo compounds t r r benzotriazoles, see K. H. Schiindehutte in "Methoden des Organischen Chemie," Vol. 10/3, Houben-WeyL Georg Thieme Verlag, Stuttgart, Germany, 1965, pp 425-430.

METALCOMPLEXES OF Azo DYES 1335

Vol. 8, No. 6, June 1969 o-Amino-0'-halodiarylazo Compounds

Formazans

Because of the just mentioned oxidative cyclization, l 4 i t was anticipated that benzotriazoles would result from the reaction of o-amino-0'-halodiarylazo compounds with ethylenediamine and cupric chloride. I n fact, the cyclization of the (0-halopheny1azo)naphthylamines XVIIIa-d to the naphthotriazoles XIXa-d has been reported to occur with cupric sulfate in hot, aqueous pyridine.15

An interesting application of the replacement of ohalogen by ethylenediamine was found in the formazan series. Unsubstituted l ,5-diphenylformazans are known2 to form 2 : 1 complexes with divalent metals and are therefore analogous to o-hydroxyazobenzenes. It was therefore reasonable to examine the reaction of a halo-monosubstituted formazan with ethylenediamine and cupric chloride. The red bromo compound X X I reacted readily at 60" or more slowly a t room temperature to form a deep blue copper complex in which the bromine was replaced.

X XVIIIa, X = F b, = c1 c,X=Br

x

d,X-I

X XIXa, X = F b, X = C1

c,X=Br d,X=I

I

C6Hj XXI

CuZc,HZNCHzCHaNHz

(X = Br, I)

Elemental analysis indicated that the product was not the simple quadridentate complex X X I I but rather that i t contained an additional 0.5 mol of ethylenediamine. It is possible therefore that the copper in

n NH2

XX

Nevertheless, when XVIIIc or XVIIId (X = Br or I) was allowed to react with cupric chloride in aqueous ethylenediamine a t 85-90', an orange solid was obtained which contained neither bromine16 nor copper. Elemental analysis suggested that the product was 1-phenylazo-2-naphthylamine (XX), which was confirmed by comparison with an authentic sample of X X . T h a t cupric ion plays no role in this surprising hydrogenolysis reaction was shown by the formation of X X in similar yield when the cupric chloride was omitted. This result indicates that ethylenediamine alone is responsible for the reduction (barring the presence of an unsuspected impurity in the ethylenediamine). At present only speculative mechanisms can be presented for the formation of X X . The explanation for the failure of XVIIIc and XVIIId, as well as X X , to be cyclized to the naphthotriazoles by cupric chloride in ethylenediamine is presumably that the cupric ion is tightly complexed by the ethylenediamine." The cupric ion is therefore unable to interact with the relatively weakly complexing oaminoazo compounds.

(15) V. Chmatal, J. Poskocil, and 2. J. Allan, Collection Czech, Chem. Commzrn.,24,484 (1959). (16) Reaction of the chloro analog XVIIIb with sodium butoxide and copper acetate in a mixture of 1-butanol and toluene at 100' for 8 hr was reported to result in replacement of 21 % of the chlorine as chloride ion. The organic product, however, was not identified: M. A. Andreeva and B. I. Stepanov, J. Gen. Chem. U S S R , 30, 1380 (1960). (17) The reaction of XVIIIa with ethanolamine is the subject of a manuscript in preparation.

XXII

X X I I is actually five-coordinate and that one ethylenediamine molecule serves to link two units of X X I I together through its amino groups. The infrared spectrum of this ethylenediamine-containing complex (XXIIa) shows only a single weak absorption in the N-H stretch region a t 3200 cm-l. When water is added to a solution of X X I I a in pyridine, a new complex X X I I b is obtained which has lost the ethylenediamine-linking molecule. Elemental analysis of X X I I b is in agreement with structure X X I I plus a molecule of pyridine and 0.5 mol of water. The presence of pyridine in X X I I b supports the fivecoordinate nature of copper in X X I I a and was confirmed by a strong peak a t m/e 79 in the mass spectrum. The infrared spectrum of X X I I b shows a weak doublet a t 3330 and 3190 cm-l; however, no bands due t o the 0.5 mol of water could be detected. Both X X I I a and X X I I b show a strong peak a t 630 nm in dimethylformamide solution. Role of the Copper Ion There arises the question of the importance of the metal ion in the displacement of the o-halogen by the amino compounds. The Hammett constant for the phenylazo group is similar (0.64) t o that for the cyano group (0.63) but

1334 THEODOR A. LISS AND DONALD R. BAER somewhat lower than that for the nitro group (0.78).18 Both of these latter groups are recognized as being strongly activating for nucleophilic displacement of ortho- or para-situated halogen atoms. This activating influence helps to explain the reactivity of halo compounds Ia, IC, and X I with ethylenediamine in the absence of metal ion. Stepanov6 has suggested that in the 2: 1 azo dyecopper complex which is probably present as an intermediate in the copper-catalyzed reactions of o-halo-ofhydroxyazo compounds, the copper interacts with the halogen so as to weaken the carbon-halogen bond and facilitate displacement of the halogen atom. The labillty of the o-halogen is also increased by the partial positive charge generated on the azo nitrogen in the 2 : 1 complex. Such effects are probably also operative in

the present reactions with alkanediamines and ethanolamine, since we observed qualitatively in a number of the reactions with ethylenediamine, and especially in the reaction with ethanolamine, that the rate was higher in the presence of cupric ion than in its absence. Kinetic studies ivi11 be needed to establish the magnitude of these effects. In a t least one instance, however, namely, the conversion of the pyrazolinone compound X I to XI1 in the presence of cupric ion, it is apparent that cupric ion does not participate. Since quadridentate XI1 does not form a complex with cupric ion in ethylenediamine solution and since bidentate X I is undoubtedly a weaker complexing agent than quadridentate X I I , an intermediate copper complex of X I cannot play a significant role in the formation of XII.19920 o-Hydroxy-o '-(methoxycarbonyl) azobenzene The concept of allowing a bidentate azo compound having a labile group to react with ethylenediamine to form a quadridentate complex was extended to an o-hydroxy-o - (methoxycarbonyl) azobenzene. Thus, when the ester X X I I I was treated with cupric chloride and aqueous ethylenediamine a t 70-80°, it was possible to isolate after purification the complex XXIV. The visible absorption maximum underwent a marked (18) H. H. JaffB, C h e w Rea., 63, 191 (1953). (19) I t is recognized t h a t even in the reactions of o-hydroxyphenylazo compounds, only a small portion of the cupric ion may actually be complexed by the starting azo compound, because the major portion is complexed by t h e alkanediamine. Stability constants which have been reported for cupric ion with ethylenediamine (KI = 1 1 2 0 " ) and P-phenylazo-~-cresol (KI = 9.671Ob) i n 75% dioxane support this view, although i t is clearly not possible t o extrapolate these values from 7570 dioxane t o aqueous ethylenediamine. I n any case the acceleration by cupric ion is then all the more remarkable. In the case of the more weakly complexing ethanolamine, a larger portion of the cupric ion will be complexed by the starting azo compound. It is thus not surprising t h a t the cupric ion caused a large rate increase in t h e ethanolamine reaction. (20) (a) L. G. Uitert and W. C. Fernelius, J . A m . Chem. Soc., 7 6 , 375 (1954); (b) F. A . Snavely, as cited in "Stability Constants of Metal-Ion Complexes," Special Publication No. 17, T h e Chemical Society, London, 1964, p 680.

inorganic Chemistry

CH, I

I

COOCH, XXIlI

Hb

,Cur0

XXIV

bathochromic shift from 401 nm ( E 6700) to 503 nm (E 11,200) (both in dimethylformamide). The presence of the coordinated primary amino group is confirmed by the weak doublet in the infrared spectrum a t 3360 and 3230 cm-l. I n the carbonyl region the band a t highest frequency appears a t 1620 cm-l (weak to medium intensity). It is likely that this band represents aromatic C=C stretching vibration. A band of strong intensity appears a t 1590 cm-l and this band is tentatively assigned to the amide carbonyl. I n any case it is clear that bonding of the copper to the amide nitrogen shifts the carbonyl absorption strongly to longer wavelength. As in the case of complexes of o-carboxy-of-hydroxyazobenzene, there is no ambiguity about which of the azo nitrogens in XXIV is bonded to metal. Since a seven-membered ring would result if the other azo nitrogen 11-ere involved, structure XXIV is much more likely. Magnetic Susceptibility The magnetic susceptibilities of a number of the complexes described above were determined. The nickel complex XVI was found to be diamagnetic, which indicates that the complex has a square-planar geometry around the nickel atom. All of the copper complexes examined (including 11, VI, XV, and XXIV) had magnetic moments equal to or greater than the spin-only value, pLeff= 1.73. The magnetic moments for the copper complexes establish the oxidation state Cu(I1) d 9 but indicate very little about the geometry of the complexes, since the moment for this state is almost independent of stereochemistry. The results do suggest that no significant interaction between copper atoms is occurring. X-Ray examination would be required to establish the stereochemistry of the copper complexes with certainty. Experimental Section All melting points are uncorrected. Infrarcd spectra we1 e determined with Perkin-Elmer infrared spectrophotometers, Models 621 and 221, and were taken as Nujol mulls unless otherwise indicated. The electronic spectra were measured with a Cary Model 14 recording spectrophotometer. The magnetic susceptibilities mere determined on a Faraday balance which was calibrated with HgCo(SCN),.

METALCOMPLEXES OF Azo DYES 1335

Vol. 8, No. 6, June 1969 The o-halophenylazo compounds in Table I (except compounds XVIIIc and XVIIId) were prepared by coupling the diazotized haloamine t o the appropriate coupling component under alkaline conditions. XVIIIc and XVIIId were prepared by coupling to 2-amino-1-naphthalenesulfonicacid under acid conditions. The solvents used for recrystallization are listed in Table I ; a Soxhlet apparatus was used in some cases.21 Reaction of o-Halophenylazo Compounds with Ethylenediamine and Cupric Chloride. (a) 1b.-To a solution of 2.5 g of I b in 25 ml of ethylenediamine was added a solution of 1.75 g of CuClz.2Hz0 in 5 ml of water plus 10 ml of ethylenediamine. The mixture was warmed to 85-90" for about 3 hr and then cooled to room temperature. The needles of I1 which formed were collected and washed with water, yielding 1.45 g. A second crop wae obtained by adding an equal volume of water to the filtrate a t 85O. The crystals were collected a t room temperature, yielding 1.6 g; total yield, 3.05 g (90%). The analytical sample of I1 was obtained by recrystallizing the crude product from methanol (Soxhlet) .21 (b) Ia and Id.-The fluoro and iodo analogs Ia and Id, respectively, were similarly treated with ethylenediamine and CuClz.2H20. The yield of I1 isolated from the fluoro analog after 4-5 hr a t 70-80' was 72%, whereas the yield was 70% after 7 months a t room temperature. Shorter reaction periods a t room temperature were not tried. The iodo analog gave a 75y0 yield (no attempt was made to isolate a second crop) of I1 after 4 hr a t room temperature. (c) Additional o-Halophenylazo Compounds.-Similar concentrations and reaction conditions were used for reaction of 111, VII, and I X with ethylenediamine, and for the reaction of IC with 1,2-propanediamine. The details of these reactions are listed in Table 111. iYo attempt was made to obtain maximum yields.

Coinplex

Rcact ion temp, " C

TABLE I11 Yo yield Time, hr

(crude)

Recrys solvent

TV 1'111 X

50 90 90

5 4.5 10

73 60 27

V

60

1

64

Methanol21 Acetonez1 (1) Methanolz1 (2) Ethylenediamine-water Acetonez1

Reaction of ICwith 1,3-Propanediamine and Cupric Chloride.To a solution of 2.9 g of IC in 40 ml of 1,3-propanediamine was added a solution of 2.5 g of CuClz.2HzO in 5 ml of water plus 15 ml of 1,3-propanediamine. The mixture was heated at 6575" for 5 hr, after which an addition of 2.5 g of cupric chloride in 5 ml of water was made. Heating was continued for about 16 hr, and the needles of VI were then collected at room temperature; yield, 1.75 g (51%). Purification was effected by twice dissolving the solid in 1,3-propanediamine and then precipitating the solid by the dropwise addition of water. Reaction of XI with Ethylenediamine. (a) In Presence of Cupric Chloride.-To a solution of 3.6 g of X I in 25 ml of ethylenediamine was added a solution of 2.5 g of CuC12*2HzOin 5 ml of water plus 10 ml of ethylenediamine. The mixture was heated a t 75-80' for 12 hr, after which 65 ml of water was added by drops a t room temperature. The yellow solid which formed was filtered off to give 2.9 g (86%) of X I I ; mp 165-169' dec; A,%,""" 425 nm (E 16,100). (b) In Absence of Cupric Chloride.-A solution of 3.6 g of X I in 25 ml of ethylenediamine was heated under nitrogen a t 75-80' for about 22 hr. The mixture was cooled and 25 ml of water was added by drops a t 40' or less. The yellow solid which formed was filtered off and washed with cold methanol and then with water; yield, 3.1 g (92%). After recrystallization from (21) T h e recrystallization was effected by extracting the solid with the indicated solvent i n a Soxhlet apparatus; the crystals which formed in the receiver were then collected.

pyridine, the yellow crystals of XI1 melted a t 166-168' dec (inserted a t 162') and gave the analysis shown below; k%'" 425 nm (E 16,000). The infrared spectra of both samples of X I I , prepared in the presence and absence of CuC12.2Hz0, were iden~ O64.3; : H , 5.99; N, 25.0. tical. Anal. Calcdfor C I ~ H Z O N ~ C, Found: C,64.3; H,6.1; N,24.7. Reaction of XI1 with Cupric Chloride.-To a solution of 1.9 g of XI1 in 50 ml of ethanolamine was added a solution of 0.8 g of CuClz.2HzO in 10 ml of ethanolamine plus 1 ml of water. After 15 min, 15 ml of water was added by drops. The rods which separated were filtered off and washed with methanol and then with water to give 1.06 g (90y0) of X I I I ; mp 197-200". Formation of X1V.-A solution of 11.6 g of IC in 120 ml of ethylenediamine was heated for 22 hr at 80-85", after which 60 ml of water was added by drons a t room temperature. The solid which formed was filtered off and recrystallized from methanol to give 4.3 g (40%) of XIV, mp 130-132"; 467 nm (E 13,700). The thin layer chromatogram (Silica Gel G, 4 : l ethyl acetate-isopropyl alcohol) showed a single orange spot. Anal. Calcd for CljH18N40: C, 66.7; H , 6.71; N, 20.7. Found: C,66.7; H, 7.0; N, 20.4. When the fluoro compound Ia was allowed to react under similar conditions (26 hr a t 75'), XIV was obtained in 79y0 yield. Reaction of XIV with Cupric Chloride.-To a solution of 0.54 g (0.002 mol) of XIV in 8 ml of pyridine a t room temperature was added by drops a solution of 0.55 g (0.0022 mol) of CuSOa.5Hz0 in 3 ml of water plus 3 ml of pyridine. After 3 nil more of water was added, the crystals of I1 were filtered off; yield, 0.44 g (66%). This product was identical with I1 prepared from Ib in one step as described above. Reaction of IC with Ethanolamine and Cupric Chloride.T o a slurry of 2.9 g of the azo compound IC in 60 ml of ethanolamine was added a t room temperature a solution of 2.0 g of CuCl2.2HzO in a mixture of 4 ml of water and 10 ml of ethanolamine. The mixture soon became homogeneous and blue-red. The mixture was stirred a t room temperature overnight, after which the red crystals of crude XV were filtered off; yield, 3.4 g. The analytical sample was obtained by twice recrystallizing the crude sample from acetone (Soxhlet).21 Reaction of IC with Ethylenediamine and Nickel Chloride.To a solution of 2.9 g of the azo compound ICin 25 ml of ethylenediamine a t 75' was added a warm solution (75') of 2.5 g of XiC12. 6Hz0 in 5 ml of water and 10 ml of ethylenediamine. After 1 hr 12 ml of water was added and the temperature was raised to 85-95' for several hours. The needles of XVI were filtered off a t room temperature; yield, 2.2 g (67%). The analytical sample was twice recrystallized from acetone (Soxhlet).21 The infrared spectrum (Nujol) shows a doublet of weak to medium intensity a t 3230 and 3110 cm-l. Reaction of ICwith Ammonia and Cupric Chloride.--A mixture of 5.8 g of IC, 5 g of CuClz.2H20, and 10 ml of concentrated ammonium hydroxide in 80 ml of formamide plus 6 ml of water was heated in an autoclave for 22 hr a t 170". A portion of the solid obtained from this reaction mixture was washed with methanol and then treated with hot isopropyl alcohol. The filtrate was concentrated on a hot plate, after which there was isolated in the cold 0.18 g of XVII (mp 129-132'), identified by comparison with the infrared spectrum and melting point (131-132') of an authentic sample of XVII. Formation of 1-Phenylazo-2-naphthylamine (XX) from XVIIIc and XVII1d.-To a solution of 3.3 g of XVIIIc in 25 ml of ethylenediamine was added a solution of 3.0 g of CuC12.2HzO in 7 ml of water and 5 ml of ethylenediamine. The resulting solution was heated at 85-90' for about 24 hr, after which 25 ml of water was added by drops a t room temperature. The semisolid which separated from solution was filtered off, washed with water, and then dissolved in benzene. This dried (Na2S04)benzene solution was then chromatographed on a column of silicic acid which contained 10% water by weight. The orange fraction which was eluted with benzene was collected, and the benzene was removed under vacuum t o give 1.5 g (60%) of crude XX, mp 87-92".

R . J. KOWATT,AND K , E.MILLER 1336 K. L. LUNDBERG, Recrystallization from cyclohexane afforded 0.93 g of orange 443 nm ( E needles, mp 99-100"; (lit.22mp 102-104"); X$f:ox 14,000). The same product was obtained from XVIIId. When the cupric chloride mas omitted from the above reaction with XVIIId, a 62% yield of XX was isolated. An authentic sample of X X was prepared by coupling phenyldiazonium chloride t o 2-naphthylamine. The two samples had the same Rp value on thin layer chromatography (Silica Gel G, 2 : 1 benzene-cyclohexane) and had very similar visible and infrared spectra. 1-(o-Bromophenyl)-3,5-diphenylformazan (XXI).-To a mixture of 12.9 g (0.075 mol) of o-bromoaniline and 22 ml of concentrated hydrochloric acid in 75 ml of water a t 5" was added by drops a solution of 5.2 g of sodium nitrite in 15 ml of water. The excess nitrous acid was then destroyed by adding a small amount of sulfamic acid. The coupler solution was made by adding a solution of 15 g (0.076 mol) of benzaldehyde phenylhydrazone in 150 ml of pyridine to a solution of 15 g of sodium hydroxide in 375 ml of methanol. The cold diazonium solution was then added by drops to the coupler solution a t 5-8". Stirring was continued a t 3-5" for 45 min, after which the crystals of X X I which had formed were filtered off and washed in turn with methanol and water; yield 13.8 g; mp 133-138". After a recrystallization from isopropyl alcohol the solid melted a t 155-156'; 486 nm ( E 14,600). Anal. Calcd for yield 8 g (28%); ":A: CIQHl6Br;\'4: C, 60.2; H , 3.99; X, 14.8; Br, 21.1. Found: C, 60.2; H,4.0; N, 14.7; Br, 20.5. Reaction of XXI with Ethylenediamine and Cupric Chloride.To the red solution of 2.5 g (0.0066 mol) of X X I a t 60" was added a solution of 1.2 g (0.007 mol) of CuC12.2H2O in 3 ml of water and 10 ml of ethylenediamine. The mixture was deep blue after 10 min a t 60" and mas then cooled t o room temperature and stirred overnight a t room temperature. The blue diamondshaped plates of XXIIa which had formed were filtered off and washed with methanol; yield, 1.2 g (42y0); mp 220-222' dec; X%LF 620 nm ( E 17,400), 500 (4500). When a similar reaction was carried out entirely a t room temperature over a 22-hr period, the yield of complex XXIIa was 627,. Anal. Calcd for (221H ~ ~ N ~ C U . O . ~ CC, ~ H58.7; ~ S ~H: , 5.37; S , 21.8; CU, 14.1. Found: C, 58.6; H , 5.3; N, 21.6; Cu, 13.7. To a solution of 0.2 g of XXIIa in 30 ml of pyridine was added

(22) T. A. Lawson, B e y . , 18, 798 (1885).

Inorganic Chemistry

15 ml of water by drops. The crystals of XXIIb which separated were collected and washed with water; yield, 0.09 g; mp 200204 dec; Fz:X 620 nm ( E 21,300), 500 (5600). Anal. Calcd C,~ 61.5; . ~ . ~H H, Z5.16; O : N, 19.3. for C Z ~ H ~ O ~ G C U . C ~ H ~ Found: C,61.5; H, 5.2; N,19.6. Coupling of Methyl Anthranilate to p-Cresol.-To a solution of 15.1 g of methyl anthranilate in 200 ml of water plus 20 ml of concentrated hydrochloric acid a t 0-5" was added slowly 20 ml of 5 N sodium nitrite. After 15 min the excess nitrous acid was destroyed by adding sulfamic acid. This diazonium solution was then added slowly a t 0-5' to a mixture of 16 g of @-cresol,5.6 g of sodium hydroxide, and 200 ml of water. The product turned crystalline near the end of the addition of the diazonium solution. The pH was maintained a t 10-11 by adding 307, sodium hydroxide until the test for diazo was negative. The yield of crude X X I I I was 23.6 g (87%); after recrystallization from methanol, the yield was 19.9 g (73%); mp 63-64". The infrared spectrum shows a strong band a t 1727 cm-I (C=O stretch). Anal. Calcd for CljH14N203: C, 66.6; H , 5.22; N, 10.4. Found: C, 66.7; H , 5.3; X, 10.4. Reaction of XXIII with Ethylenediamine and Cupric Chloride. --A solution of 5.1 g of CuC12.2HzO in 10 ml of water mas added slowly to a solution of 5.4 g of X X I I I in 40 ml of ethylenediamine. The mixture was then heated overnight a t 70-80", after which 60 ml of water was added by drops a t room temperature. The resulting solid was filtered off, washed with a 2 : 1 water-ethylenediamine mixture, followed by water; yield, 6.1 g. A 2-g portion of this solid was extracted vith 125 ml of hot methanol and the resulting solution was chromatographed on a silicic acid (5% water content) column, using methanol as the eluent. The main orange fraction was collected and concentrated t o dryness to give 0.88 g of XXIV. The product gave a single spot on thin layer chromatography and was recrystallized (Soxhlet)21from methanol before analysis; yield, 0.59 g. Anal. Calcd for CIGHI&~O~CU: C, 53.4; H,4.5; N, 15.6. Found: C, 53.4; H,4.6; N, 15.4.

Acknowledgment.--We thank Dr. Robert K. Miller for his assistance in the interpretation of the infrared spectra, Dr. James F. Weiher for determining and interpreting the magnetic susceptibility data, and Bernard Blinchikoff and Charles R. Haff, Jr., for skillful assistance in the laboratory.

CONTRIBUTION FROM THE CHEMISTRY DEPARTMENT, 57069 UNIVERSITY OF SOUTH DAKOTA, T T ~SOUTH ~ DAKOTA ~ ~ ~

gem-Dibasic Ligands with Phosphorus, Sulfur, and Nitrogen Sites, and Some Borane Derivatives: BY K. L. LUXDBERG,2&R. J. ROWATT,2a A N D K. E. MILLERZb

Received Novenzbev 11, 1968 Three gem-dibasic ligands, dimethylphosphinomethyldimethylaniine [( CHa),PCH,S( CHI)Z] (pcn), dimethylphosphinomethyl methyl sulfide [( CH3)2PCHnSCH3](pcs), and dimethylaminomethyl methyl sulfide [( CH3)&CHzSCH3] (ncs) were prepared using chloromethyldimethylamine or chloromethyl methyl sulfide as starting materials. All three are colorless liquids of low vapor pressure (5-10 mm a t room temperature), pcn and pcs being very inflammable in air. Monoborane adducts, HsBpcn, H3Bpcs, and HsBncs, form readily with diborane. The only stable diadduct was H3BpcnBH3. A new borane cation type was demonstrated in the preparation of CHapcnBHa+.

The preparation and characterization of some 1,3dihetero skeletons mere conducted as a part of a syn(1) Presented a t t h e h l i d a e s t Regional IIeeting of t h e Amelican Chemical Society, Manhattan, Kan., Oct 31, 1968. (2) (a) National Defense and Education Act Fellows. (b) Author t o whom correspondence should be directed a t t h e University of South Dakota.

thetic program in small ring chemistry. Of special interest in this program were compounds with two differentheteroatom base sites that could be used to incorporate functional chemistry into complexes without the Complication Of chelate formation Or polymerization,

~

~